Chart Review and Questionnaire
Between March 1987 and July 1995, 183 triple arthrodeses were performed in 160 patients at our institution by one of the two senior ones of us (L. C. S. or M. S. M.). All of the patients had an arthrodesis of the talonavicular, calcaneocuboid, and talocalcaneal joints with rigid internal fixation with use of partially threaded 6.5-millimeter cancellous-bone screws. Of the 160 patients, forty-nine were excluded: thirty-four were lost to follow-up, seven were unable or refused to participate, six had moved away, and two had died. The remaining 111 patients (132 feet) were followed for an average of 5.7 years (range, 2.0 to 10.8 years) and formed the study group. The average age of these forty-five male patients and sixty-six female patients at the time of the operation was 54.9 years (range, thirteen to seventy-seven years). Pertinent coexisting medical conditions included heart disease (twelve patients), osteoarthritis (twelve patients), diabetes (nine patients), endocrinological disease (six patients), gout (four patients), connective-tissue disease (three patients), psoriasis (three patients), and lupus (one patient).
Each patient completed a questionnaire that included the subjective portion of the modified ankle-hindfoot scale of the American Orthopaedic Foot and Ankle Society17 and certain questions pertaining to function and outcome (Table I). This 100-point scale was modified to a maximum possible score of 94 points by subtracting the 6 points assigned to subtalar motion, which was expected to be absent in this study population. Additionally, the patients were asked to rate their satisfaction with the procedure on a scale (not a visual analog) of 0 (not satisfied) to 10 (completely satisfied without reservation) and to state whether they would have the operation again under similar circumstances.
The 132 ankles and feet were divided into one of five preoperative diagnostic groups (posttraumatic changes, rheumatoid arthritis, rupture of the posterior tibial tendon or flatfoot, neuromuscular imbalance, or miscellaneous) and one of three preoperative deformity groups (valgus deformity, varus deformity, or neutral position) (Table II).
Physical Examination
Follow-up physical examination included the objective portion of the ankle-hindfoot scale17, excluding the 6 points for subtalar motion. All of the patients were examined by one of us.
Radiographic Assessment
Postoperative anteroposterior and lateral radiographs of the involved foot and ankle were made while the patient was bearing weight. The talus-first metatarsal angle and the talonavicular coverage angle were measured on the anteroposterior radiograph, and the talus-first metatarsal angle was also measured on the lateral radiograph. All of the angles were measured with the method described by Sangeorzan et al.23. Degenerative joint disease of the ankle was assessed on anteroposterior and lateral radiographs, made with the patient bearing weight, according to a grading scale (Table III) developed as a combination and modification of two existing scales5,13. The scale of Graves et al.13 (four grades) is applicable to any unfused joint in the foot (not exclusively the ankle joint), and the scale of Bauer et al.5 (four grades) offers less detail than our scale. In our study, for patients who had valgus or varus deformity of the ankle, the side of the joint that had a decrease in the joint space was used for measurements.
The same parameters were measured on the preoperative radiographs, and the preoperative and postoperative measurements were compared. All radiographic analyses were performed by one of us (R. F. P., IV)
Operative Technique
After the induction of anesthesia, a tourniquet was placed around the proximal part of the thigh of the involved limb. The leg was prepared, draped in the usual sterile fashion, and then elevated and exsanguinated. The tourniquet was inflated to 350 millimeters of mercury (46.66 kilopascals).
To determine the need for first performing a lengthening of the Achilles tendon for the treatment of equinus or severe planovalgus deformity with contracture of the Achilles tendon, the subtalar joint was held in neutral position and the ankle was dorsiflexed with the knee in 30 degrees of flexion. If the foot remained in more than 5 degrees of equinus, the Achilles tendon was lengthened; this was performed in forty legs. One lateral transverse and two medial incisions were made directly over the tendon, and the tendon was stretched, improving dorsiflexion of the foot.
Attention was then directed to the lateral aspect of the foot. A long, curved incision was made, beginning at the peroneal tendons, posterior and lateral, and curving over the dorsolateral aspect of the foot toward the extensor digitorum communis tendon dorsolaterally. The incision, which was dorsal to the path of the sural nerve, was then deepened through the subcutaneous tissue. Dorsally, the extensor digitorum longus tendon, the peroneus tertius tendon, and the anterior process of the calcaneus were identified. The incision was deepened through the subcutaneous tissue to the extensor digitorum brevis muscle sheath. The sheath was incised, and the extensor digitorum brevis muscle was reflected distally off its origin on the calcaneus and the lateral wall of the talus, exposing the anterior process of the calcaneus.
The bifurcate ligament was removed, soft tissue was removed from the sinus tarsi with a curet, and the anterior process of the calcaneus was debrided. The calcaneocuboid joint was identified and opened by adducting the forefoot to better view the joint. With use of an elevator, the interosseous ligament was stripped from the attachment to the calcaneus and the joint was denuded of cartilage. A lamina spreader was used to expose the entire posterior facet and subtalar joint, and a curet and a sharp chisel were used to denude it of cartilage. The lateral aspect of the talonavicular joint was similarly approached with a rongeur and a curet. A second longitudinal incision, medial to the anterior tibial tendon, was made over the talonavicular joint dorsomedially; the joint was denuded of cartilage with a curet and a rongeur.
When it was determined intraoperatively, by direct inspection or fluoroscopic examination, that there was a gap at the site of the arthrodesis, bone graft was used (forty patients, fifty-three procedures). As a matter of operative routine, the first twenty-six patients (thirty-nine procedures) were managed with bone graft from the iliac crest (thirty-eight procedures) or the greater trochanter (one procedure); then, after a review indicated that donor-site problems could be avoided with the use of local bone, the remaining fourteen patients (fourteen procedures) were managed with local bone graft.
The iliac crest bone graft was obtained in standard fashion with a long incision over the anterior superior iliac spine, with avoidance of the anterior spine. The incision was deepened through subcutaneous tissue, and the periosteum was incised with electrocautery. The periosteal flap was raised, and a copious amount of cancellous bone was removed. The iliac crest donor site was closed over a small Hemovac drain with interrupted 0 Vicryl sutures (polyglactin; Ethicon, Somerville, New Jersey) for the periosteum, followed by interrupted 2-0 Vicryl sutures for the subcutaneous tissue and 4-0 nylon sutures for the skin. The bone graft was packed into the subtalar, calcaneocuboid, and talonavicular joints. To obtain bone graft from the greater trochanter, a longitudinal incision was made over the greater trochanter and deepened through the subcutaneous tissue down to the trochanter, the soft tissue was retracted, and a periosteal window was made. A copious amount of cancellous bone chips was removed with a curet. The periosteal window was closed with 2-0 Vicryl sutures, and a layered closure was then performed with 4-0 Vicryl sutures followed by 4-0 nylon sutures. Local bone graft was obtained from the talus and the calcaneus and was packed into the sites of the calcaneocuboid, talonavicular, and subtalar arthrodeses.
Six feet and ankles had additional osseous procedures, which included four calcaneal osteotomies, one partial fibulectomy, and one resection of a coalition. The calcaneal osteotomies were performed in the presence of residual calcaneal varus or valgus angulation after performing the arthrodesis. The partial fibulectomy was performed in a patient who had fibular-calcaneal abutment with subsequent development of an osteophyte preoperatively. The coalition was resected to permit mobility between the bones and, thus, to obtain a better anatomical alignment.
The talonavicular joint was reduced and the hindfoot was brought into neutral position, allowing excellent alignment of the osseous surfaces of the subtalar, talonavicular, and calcaneocuboid joints. In this position, the talonavicular joint was transfixed with a guide-pin, and the subtalar joint was fixed with a guide-pin from the tip of the heel to the talus. Either 6.5 or 7.0-millimeter cannulated, large-fragment, distally threaded AO screws were inserted from the calcaneus into the talus, from the navicular into the talus, and from the calcaneus into the cuboid (Fig. 1-A, 1-B, 1-C, 1-D, 1-E, 1-F).
The alignment of the foot was checked radiographically after temporary fixation. Although we do not document intraoperative changes of fixation, we now routinely use cannulated screws for fixation and more than 50 percent of feet need a change in the position of the guide-wire intraoperatively. All of the wounds in this area were irrigated with normal saline solution and were closed with 2-0 Vicryl sutures; 4-0 nylon sutures were used for the skin. After deflation and removal of the tourniquet, sterile dressings and a posterior splint were applied to hold the foot and ankle in proper position.
Postoperative Management
The posterior splint was removed after ten days. The patient was non-weight-bearing while wearing a below-the-knee cast for four weeks, and then a weight-bearing cast was used for an additional six weeks until osseous union was evident radiographically.
Analyses
The chi-square test was used to evaluate comparative results. The level of significance was defined as p = 0.05.
Subjective Assessment
Overall, patient satisfaction averaged 8.3 points (range, 0 to 10 points) on the 0 to 10-point grading scale, and 101 (91 percent) of the 111 patients stated that they would have the procedure again under similar circumstances. Of the twenty-one patients who had a bilateral operation, two (10 percent) stated that they would not have the operation again. With the numbers available for study, we could not detect a significant difference in terms of satisfaction or willingness to repeat the procedure among the five diagnostic groups or among the three deformity groups (Table IV).
Eighteen (16 percent) of the 111 patients were able to wear fashionable shoes at the time of follow-up, seventy-one (64 percent) could wear comfortable shoes only, and twenty-two (20 percent) needed modified shoes or an ankle-foot orthosis. With the numbers available for study, we could not detect a significant difference in terms of types of footwear among the five diagnostic groups or among the three deformity groups.
With the numbers available for study, we found no significant difference in terms of patient satisfaction or the type of footwear between the forty feet that had been treated with concurrent lengthening of the Achilles tendon and the ninety-two that had been treated without concurrent lengthening. The result was considered satisfactory for thirty-seven of the forty feet that had had the lengthening and for eighty-five of the ninety-two feet that had not.
Clinical Assessment
The average ankle-hindfoot score for all 132 feet was 60.7 points (range, 0 to 94 points). With the numbers available for study, we could not detect a significant difference in this score among the five diagnostic groups or among the three deformity groups (Table IV).
There was a significant association between satisfaction and hindfoot alignment (p = 0.001). With the numbers available for study, we could not detect a significant difference in terms of the average ankle-hindfoot score between the forty feet treated with lengthening of the Achilles tendon (55.7 points) and the ninety-two treated without lengthening (62.2 points).
Radiographic Assessment
Measurement of Angles and Height
Comparison of the preoperative and postoperative anteroposterior radiographs demonstrated that, overall, the talus-first metatarsal angle had improved an average of 14 degrees (range, -7 to 45 degrees). The twenty-two feet affected by rheumatoid arthritis and the seventy feet with either a rupture of the posterior tibial tendon or a severe flatfoot deformity showed the most improvement. With the numbers available for study, we could not detect a significant difference in this improvement among the five diagnostic groups or among the three deformity groups.
Overall, the talonavicular coverage angle improved an average of 26 degrees (range, -8 to 60 degrees) as demonstrated by a comparison of the preoperative and postoperative anteroposterior radiographs. The twenty-five feet with posttraumatic changes needed the least correction, and the seventy feet with rupture of the posterior tibial tendon or severe flatfoot deformity needed the most (Table IV). The difference between these two groups was significant (p = 0.0022).
Evaluation of the preoperative and postoperative lateral radiographs showed that, overall, the lateral talus-first metatarsal angle had improved an average of 14 degrees (range, -17 to 41 degrees). The twenty-five feet with posttraumatic changes needed the least correction, and the twenty-two feet with rheumatoid arthritis and the seventy feet with rupture of the posterior tibial tendon or flatfoot needed the most (Table IV). The difference between the rheumatoid group and the posttraumatic group was significant (p = 0.001), as was the difference between the group with rupture of the posterior tibial tendon or flatfoot and the posttraumatic group (p = 0.001). Among the deformity groups, the 102 feet with valgus angulation needed the most correction and the seventeen feet with a neutral position needed the least (Table IV); the difference between these two groups was significant (p < 0.001).
Arthritis
Nine (7 percent) of the 132 ankles had degenerative joint disease preoperatively, and seventy-nine (60 percent) had clear signs of progressive degenerative joint disease at the time of follow-up, which is a significant difference (p < 0.01) (Figs. 2-A, 2-B, 2-C, 2-D, 2-E, 3-A, 3-B, 4). At the follow-up evaluation, grade-I arthritis was found in 28 percent (twenty-two) of the seventy-nine ankles (Figs. 2-C, 2-D, and 2-E); grade-II arthritis, in 19 percent (fifteen) (Fig. 3-B); and grade-III arthritis, in 11 percent (nine) (Fig. 4). Five ankles had had such severe degenerative joint disease that they had already been treated with an arthrodesis before presentation for this study.
With the numbers available for study, we could not detect a significant association between sagittal motion of the ankle and satisfaction of the patient or between the prevalence of radiographic evidence of arthritis of the ankle at the time of follow-up and satisfaction. However, there was a significant association between decreased range of motion of the ankle and severity of arthritis of the ankle (p = 0.01). Delineation of sagittal motion of the ankle was based on the three-tiered scale (30 degrees or more, 15 to 29 degrees, and less than 15 degrees) of the ankle-hindfoot score. The average grade of arthritis of the ankle seen radiographically for each of these tiers was 0.8 (range, 0 to 3), 1.2 (range, 0 to 3), and 2.6 (range, 1 to 3), respectively. We could not detect a significant difference among the five preoperative diagnostic groups or the three preoperative deformity groups in terms of the prevalence of radiographic evidence of degenerative joint disease of the ankle preoperatively or at the time of follow-up.
Time to Fusion
Fusion was determined clinically and radiographically. Overall, the time to fusion averaged 10.7 weeks (range, 3.4 to 24.6 weeks) (Table IV). With the numbers available for study, we could detect no significant difference among the five preoperative diagnostic groups or the three preoperative deformity groups with regard to the time to fusion.
Complications
There were eleven complications in ten patients. The complications included four superficial wound problems, three nonunions (two calcaneal and one subtalar), one case of superficial peroneal neuritis, one case of Charcot-like neuroarthropathy of the foot (in a patient in whom diabetes developed during the follow-up period), one rupture of the Achilles tendon, and one case of peroneal tenosynovitis.
We reviewed the literature extensively and found only two studies in which different preoperative diagnostic and deformity groups were related to outcome. Seitz and Carpenter25 found that residual deformity was most frequent in patients who had spastic paralysis (a prevalence of fourteen of nineteen), painful flatfoot (four of five), or preoperative valgus deformity (twenty-two of twenty-five). Of the preoperative diagnostic and deformity groups of Angus and Cowell3, the one with idiopathic clubfoot and the one with rigid equinovarus had the poorest results. In our study, all of the triple arthrodeses involved rigid internal fixation without resection of bone wedges, as compared with twelve of eighty in the study by Angus and Cowell3 and thirty of sixty-six in the study by Seitz and Carpenter25.
With the numbers available for study, we could detect no significant difference in outcome among the five diagnostic or the three deformity groups. In addition, patient satisfaction was high and most of the patients stated that they would be willing to have the procedure again under similar circumstances. To our knowledge, no previous study has documented the amount of radiographic correction needed by different groups. We found that, in general, the patients who had the procedure for the treatment of posttraumatic changes needed the least amount of correction, whereas those who had rheumatoid arthritis or had rupture of the posterior tibial tendon or flatfoot deformity needed the most correction in all planes. Of the deformity groups, the one with a valgus foot preoperatively needed the most correction in all planes.
Wetmore and Drennan27 concluded that changes of the ankle joint are inevitable after triple arthrodesis because of the loss of the shock-absorber capacity of the hindfoot. They found that the prevalence of arthritis was 77 percent (twenty-three of thirty) on radiographic examination and 63 percent (nineteen of thirty) on clinical examination at the time of follow-up. Angus and Cowell3 found degenerative changes in thirty-one (39 percent) of eighty ankles. One foot in their series was later amputated, and, of the remaining seventy-nine, nineteen were substantially painful at the time of follow-up; thirteen of the nineteen painful feet had degenerative changes. Southwell and Sherman26 noted, on radiographs, a 58 percent prevalence of degenerative joint disease of the ankle (twenty-one of thirty-six) after triple arthrodesis, but clinically the ankles were usually asymptomatic.
In contrast, Robins22 concluded that "fixation of a given joint in the lower extremity does not have a harmful effect on the one above, especially the ankle in relation to the foot." Duncan and Lovell11 did not find a high prevalence of arthritis of the ankle after triple arthrodesis. Most recently, de Heus et al.9 reviewed the outcomes of thirty-seven triple arthrodeses after an average duration of follow-up of ten years. They concluded that arthritis of the tibiotalar joint, although a possibility, did not occur with great frequency following triple arthrodesis.
The prevalence of radiographic evidence of degenerative joint disease in our study was consistent with the findings of Southwell and Sherman26. We also observed that these radiographic findings were associated with a high prevalence of clinical degenerative joint disease of the ankle, as evidenced by a decreased range of motion of the ankle. We agree with other authors3,6,9,12,13,27 that the cause of arthritis after triple arthrodesis is the increased stress placed on the ankle joint secondary to fusion of the hindfoot.
However, arthritis of the ankle was not related to patient satisfaction. In other words, the patients were pleased despite clinical and radiographic evidence of degenerative joint disease of the ankle. Graves et al.13 found no association between the preoperative diagnosis and the severity of degenerative joint disease of the ankle after triple arthrodesis. With the numbers available for study, we could not detect an association between preoperative diagnosis or deformity and the prevalence of degenerative joint disease of the ankle preoperatively or at the time of follow-up.
In 1958, Banks and Green4 advocated heel-cord lengthening in combination with triple arthrodesis to correct a severe deformity of the hindfoot. Specifically, they stated that heel-cord lengthening must be performed in combination with a triple arthrodesis because of the amount of bone that would otherwise need to be resected to establish a plantigrade foot when treating severe equinus deformity. Herold and Torok15 recommended a two-stage operation that is initiated with lengthening of the Achilles tendon. The triple arthrodesis could then be performed with resection of a minimum amount of bone.
Howorth16 noted that concurrent lengthening of the Achilles tendon with triple arthrodesis may be needed to treat equinus deformity. Bernau7 lengthened the Achilles tendon at the time of twenty-six (52 percent) of fifty triple arthrodeses. Angus and Cowell3 thought that a more conservative resection of bone could be performed if a soft-tissue release was included with the triple arthrodesis. Overall, the patients with a preoperative deformity of rigid equinovarus had the poorest results in that series. However, the patients who had concurrent lengthening of the Achilles tendon did much better. Angus and Cowell3 concluded that "attempts to correct the deformity by bony resection alone may result in avascular necrosis of the talus, persistence of the deformity, poor bony contact at the midtarsal joint resulting in pseudoarthrosis, or any combination of the three." In 1989, McCluskey et al.18 recommended a triple arthrodesis with lengthening of the Achilles tendon to treat cavovarus deformity of the foot.
With the exception of Angus and Cowell3, none of the investigators mentioned earlier compared the outcome for patients who had had lengthening of the Achilles tendon with the outcome for those who had not. In our study, we could detect no significant differences in terms of the ankle-hindfoot score, patient satisfaction, or type of footwear worn postoperatively between the patients managed with a concurrent lengthening of the Achilles tendon and those managed without it.
In conclusion, triple arthrodesis is effective in relieving pain and decreasing functional deficits secondary to various etiologies and deformities. Although there was a high prevalence of subsequent clinical and radiographic evidence of arthritis of the ankle, we could detect no association between patient satisfaction and arthritis. The patients who had extreme deformity and needed concurrent lengthening of the Achilles tendon did not have a poorer outcome.